Enhancing and Validating Modern Radiosurgery for Intracranial Metastases

The recently-developed single-isocenter multi-target volumetric modulated arc therapy (VMAT) provides several clinical benefits including significantly decreased treatment time and cost-effectiveness compared to other traditional therapies for patients presenting with brain metastases. Despite its benefits and increasing popularity in clinical use, certain challenges associated with the technique demand investigation. (1) Current spatial accuracy recommendations are not designed explicitly for this technique. Also, many supplemental end-to-end tests specific to this technique have limitations that may be overcome by the use of a novel end-to-end test using 3D-dosimetry. (2) Several prior studies have raised questions regarding the ability of clinical dose calculation algorithms to accurately model the dose to targets with far distances from the isocenter. Discrepancy between planned and measured dose may be due to an inaccurate modeling of the machine’s dosimetric-leaf-gap (DLG), limitations in beam modeling, or other factors. Therefore, a comprehensive dosimetric examination specifically for targets far off-isocenter is warranted. (3) The single-isocenter multi-target VMAT technique has potential applications to preclinical small-animal irradiation. There are treatment planning, treatment geometry, and immobilization considerations that are involved with determining the feasibility of this application. The purpose of this thesis is to examine these aspects in order to enhance and validate the current state of the treatment of intracranial metastases.

In all examinations in this work, the clinical treatment planning system used was Varian External Beam Treatment Planning Software (TPS) version 13.6 (Varian Medical Systems) along with Anisotropic Analytical Algorithm (version 13.6.23) for dose calculation. All treatment plans in this work were modeled on a Varian Truebeam STX linear accelerator with HD-MLCs.

The presented end-to-end test utilized the 3D-dosimetry system using an n-isopropylacrylamide (NIPAM)-based polymer gel that has the ability to be read out using x-ray CT. This dosimeter was also used in the dosimetric evaluation in combination with the PRESAGE®/optical-CT 3D-dosimetry system. Point-dose measurements to supplement the 3D-dosimetry for the dosimetric accuracy evaluation were performed using the commercially-available Stereotactic End-to-End Verification Phantom (STEEV) manufactured by Computerized Imaging Reference Systems, Inc., (Norfolk, VA). All analysis for these examinations was performed using the Eclipse TPS, Matlab, and the 3D Slicer software. Immobilization considerations for the high-throughput small-animal study involved designing a 3D-model using Autodesk Fusion 360 and printing various iterations of designs using an Ultimaker 2 3D-printer.

In this thesis, we present a complete end-to-end test for the single-isocenter multi-target VMAT treatment planning technique using the NIPAM-based dosimeter. We also present this dosimetry system’s capability for remote dosimetry from the developing group, and methods for performing in-house dose calibration. In the dosimetric accuracy evaluation, we report an 8.36% percent difference between the mean dose ratio for the target located at the treatment isocenter compared to the farthest target in our plan (10 cm from isocenter) from 3D-dosimetry. Further, average gamma analysis for the target at isocenter was 87.8% and 97.95% for 1 mm/ 5% and 2 mm/ 3% criteria, respectively. For the most distal target, the average passing rates were 73.93% and 94.2% using the same criteria. While the largest dose discrepancies were noticed for the target farthest from the isocenter, a general trend of decreasing dose coverage for increasing target distance was present in all 3D-dosimetry trials. Our treatment planning examination of the feasibility of using this technique for preclinical small-animal irradiation shows adequate conformity and sparing for half-brain irradiation for up to six mice at a time. In this treatment plan, we report less than 10% variability among the six targets in median dose to the targeted and contralateral hemispheres.